1. Field of the Invention
The present invention relates to a bipolar junction transistor (BJT) and its fabricating method, and more particularly, to a self-aligned BJT and a self-aligned method for producing the same.
2. Description of the Prior Art
A bipolar junction transistor utilizes two type of carriers, “electrons” and “hot holes”, to transmit current. The BJT continues to be a basic circuit element in integrated circuits due to its high switching capability and current carrying capacity. However, the electrons/holes mobility of a silicon BJT, which has a silicon substrate, is lower and not suitable for being applied in high frequency devices. Therefore, the silicon BJT has been replaced with a GaAs BJT, which has a GaAs substrate, or a SiGe hetero-junction bipolar transistor (HBT) in recently semiconductor processes.
Please refer to FIG. 1 to FIG. 5, which are schematic diagrams illustrating a conventional method for forming a silicon BJT. As shown in FIG. 1, the silicon BJT is formed on a P-type single crystal silicon substrate 10 including a heavily doped N type (N+) region 12 and heavily doped P type (P+) regions 14 formed on the substrate 10. First, an N-epitaxial layer 16 is thermally grown on the substrate 10, which has a thickness of about 1.2 micrometers (μm). During formation of the epitaxial layer 16, dopants of the N+ region 12 and the P+ region 14 diffuse upward into the epitaxial layer 16. Then, a silicon dioxide layer 18 and a silicon nitride layer 20 are formed on the epitaxial layer 16, respectively. A photolithographic and etching process is performed to remove a portion of the silicon nitride layer 20, the silicon dioxide layer 18 and the epitaxial layer 16 to expose a portion of the epitaxial layer 16, thus forming a plurality of openings 22 in the silicon nitride layer 20, the silicon dioxide layer 18, and the epitaxial layer 16.
As shown in FIG. 2, a thermal growth process is performed to oxidize the exposed epitaxial layer 16 to an oxide layer 24, which fills in the opening 22, with a top surface of the oxide layer 24 being approximately equal to a top surface of silicon dioxide layer 18. After that, the remained silicon nitride layer 20 is removed. During thermal growth of the oxide layer 24, the dopants of the N+ region 12 and the P+ region 14 are diffused further upward to broaden the N+ region 12 and P+ region 14. Afterwards, a photoresist layer 26 is deposited on the oxide layer 24, and the photoresist layer 26 is exposed and developed to form an opening 27 in the photoresist layer 26. Then, suitable N-type dopants, such as phosphorus (P), are ion implanted through the opening 27 to form an N+ collector region 28 within the epitaxial layer 16 and beneath the oxide layer 24.
As shown in FIG. 3, the remnant photoresist layer 26 is removed and another patterned photoresist layer (not shown in FIG. 3) is formed on the oxide layer 24. A wet etching process or a reactive ion etching (RIE) process is performed to remove a portion of the oxide layer 24 above the epitaxial layer 16, thus forming an opening 30 in the oxide layer 24. Then, a P-polysilicon layer 32 and a silicon dioxide layer 34 are formed on the substrate 10, respectively. Typically, the polysilicon layer 32 has a thickness of about 8000 angstroms (Å), and the silicon dioxide layer 34 has a thickness of about 5000 to 6000 Å. A patterned photoresist layer (not shown in FIG. 3) is formed on the silicon oxide layer 34, and a photolithographic and etching process is performed to remove a portion of the silicon dioxide layer 34 and the polysilicon layer 32 to form an opening 36 in the silicon dioxide layer 34 and the polysilicon layer 32, and a portion of the epitaxial layer 16 is exposed.
As shown in FIG. 4, a silicon dioxide layer 38 is thermally grown on the substrate 10 that covers a bottom and a sidewall of the opening 36 uniformly. Typically, the silicon dioxide layer 38 has a thickness of between 0.2 to 0.4 μm. During formation of the silicon dioxide layer 38, P-type dopants of the polysilicon layer 32 diffuse downwards into the N-epitaxial layer 16 to form P-extrinsic base regions 40 within the epitaxial layer 16. Then, a directional RIE process is performed to remove a portion of the silicon dioxide layer 38 in the bottom of the opening 36 and on the oxide layer 24. Further, a relatively low-energy, high-dose ion implantation is performed to form an N shallow emitter region 42 within the epitaxial layer 16, which has a thickness of about 0.2 μm, by implanting the arsenic (As) ions through the opening 36 into the epitaxial layer 16. Similarly, a relatively high-energy, high-dose ion implantation is performed to form an N+ raised subcollector 44, which has a thickness of about 0.2 micrometers, by implanting phosphorus (P) ions through the opening 36 into the epitaxial layer 16. The energy and heat cycling conditions for the phosphorus implantation are selected such that the bottom of the raised subcollector 44 nominally submerges into the N+ region 12. And a relatively middle-energy, low-dose boron ion implantation is performed to form an intrinsic base region 46 beneath the shallow emitter region 42 and above the raised subcollector 44, by implanting the As ions through the opening 36 into the epitaxial layer 16.
Finally, as shown in FIG. 5, a photolithographic and etching process is performed to form two contact vias (not shown in FIG. 5) on the polysilicon layer 32 and the collector region 28. Then, a metal layer is filled in the contact vias and the opening 36 to form three metal contacts 50, 52, and 48, and the conventional BJT is completed.
As stated above, the conventional method for forming the BJT is very complicated. Although the intrinsic base region 46 and the emitter region 42 of the BJT are formed by the self-aligned method, the extrinsic base regions 40 are formed by the thermal diffusion process. So that the contact area between the intrinsic base region, the extrinsic base region and the collector region cannot be controlled precisely. Thus causing higher capacitance between the base region and the collector region and the conventional BJT is not suitable for being applied in high frequency devices. In addition, the interference of the PN junction of the BJT is clearer, the performance of the device is better. However while the SiGe HBT can function in high temperatures, the interference between the SiGe epitaxial layer and silicon of the SiGe HBT will generate a plurality of intersectional arrangements and thus destroy the its high speed characteristic. For this reason, the temperature for forming the SiGe epitaxial layer has to be controlled to be below 700° C. But the conventional BJT has to utilize the thermal diffusion process many times, and therefore the conventional method is not suitable for forming the HBT.